The USDA-NRCS, Plant Materials Center (PMC) was established in Booneville, AR in 1987 and serves the plant material needs of the Southern Ozarks, Arkansas River Valley, and Boston and Ouachita Mountains Major Land Resource Areas (MLRAs 116A, 118A, 117, and 119, respectively). The Booneville Plant Materials Center (PMC) is located along the Petit Jean River in Logan County, Arkansas at an elevation of 146 m (Burner, 2012). The PMC lies along the north edge of the Ouachita National Forest and is in the eastern portion of MLRA 118A, Arkansas Valley and Ridges. The 30-year normal for mean annual precipitation in Booneville is 127 cm and precipitation is delivered more or less evenly throughout the winter (23%), summer (21%), spring (29%), and autumn (27%) seasons (NOAA, 2010). The mean annual air
temperature for the region is 15.6 °C, with a winter minimum of 10.5 °C and a summer maximum of 32.4 °C (NOAA, 2010).
The study site was located on a Leadvale silt loam (fine-silty, siliceous, semiactive,
between 30 and 60 cm in the profile, which can impede water infiltration and plant root growth (Figure 1). The Leadvale soil is poorly drained in late winter and early spring, due to the presence of a perched water table. The fragipan also causes plants to experience drought stress during the summer by limiting the depth of plant roots seeking moisture stored deeper in the soil profile.
The study site was a pasture dominated by tall fescue prior to being prepared for the original yield study in 2006. No fertilizer was applied and the vegetation was kept mowed to a height of approximately 15 cm year-round. Soil samples were collected from throughout the 0-6 cm depth study site prior to planting. Soil samples were analyzed and nutrient levels were brought up to uniform medium production values prior to planting according to recommended values for P and K for native warm-season grass establishment.
2.2. Initial Switchgrass Field and Treatment Establishment
‘Alamo’ and ‘Cave-in-Rock’ switchgrass cultivars were planted at the Booneville PMC in 12.2 m (40 ft) x 12.2 m (40 ft) plots on 5 March 2007. Each plot contained one switchgrass cultivar and fertilizer source (Figure 3). Permanent overhead sprinklers were installed to deliver irrigation to select plots. Each 12.2 m x 12.2 m plot was divided in half, with one subplot harvested once per year and the other subplot harvested twice per year (Figure 2). There were a total of 24 plots including cultivar-irrigation-fertilizer source treatment combinations. There were a total of 48 subplots when harvest frequency (once or twice per year) was included.
Plots were seeded at a rate of 4.4 kg ha-1 (5 lb ac-1) pure live seed and planted with a no-tillage native grass drill (Sukup 2050 series, Jonesboro, Arkansas). After drilling, the seedbed
was rolled with a water-filled roller to establish good seed-to-soil-contact. Temporary sprinkler irrigation was applied to all plots for initial seed emergence and establishment. A permanent sprinkler irrigation system was installed in replicated irrigation treatments in the summer of 2007. Rain gauges were placed in the irrigated plots to calibrate the irrigation delivery system.
Irrigation-treated plots received 2.54 cm of irrigation water from a nearby pond per week during June through August from 2008 to 2011. The average annual precipitation during this study was 131 cm (Burner, 2012), which was slightly greater than the normal 30-year average of 121 cm for this area (NOAA, 2010). Annual rainfall was 127% in 2008, 124% in 2009, 70% in 2010, and 110% in 2011 of the normal 30-year average (Burner, 2012; NOAA, 2010). The average annual air temperature during this study was 15ºC (Burner, 2012), which was numerically less than the 30-yar average of 15ºC (NOAA, 2010). Average annual air temperature was 90% in 2008, 89%
in 2009, 93% in 2010, and 99% in 2011 of the normal 30-year average (Burner, 2012; NOAA, 2010).
The study area was burned each year in early March to remove residue stubble, stimulate switchgrass seed production for wildlife, and to remove surface residue for native pollinator nesting habitat and to create corridors for other ground-nesting wildlife species (USDA-NRCS Arkansas, 2009). Fertilizer treatments were applied when green foliage appeared in early April of 2008 to 2011.
Poultry litter was applied at a rate of 4.5 Mg ha-1 (2 ton ac-1), while commercial fertilizer was applied to match N, P, and K levels in the applied poultry litter. Litter batches were analyzed annually for nutrient concentrations prior to application.
Two harvest frequencies were imposed to test their effects on annual aboveground biomass production. A single harvest was made in November after the first killing frost for the
1-cut system. In the two-1-cut system, harvests occurred twice per year. The first harvest occurred in June just prior to the boot stage (when seedhead emerges from switchgrass), and the second harvest occurred after a killing frost in November.
2.3. Soil, Plant and Hydrologic Property Measurements
Soil samples were collected in all plots from the 0-10 and 10-20 cm depth interval for bulk density, Mehlich-3 extractable soil nutrients, soil organic matter, and soil particle-size determinations. In the top 10 cm, bulk density samples were collected manually with a 5.0-cm diameter, stainless steel core chamber and a slide hammer. For the 10-20 cm depth, bulk density samples were collected with a 5-cm diameter, stainless steel hydraulic probe. Samples from both depths were dried in a forced-air dryer at 70°C for 48 hours, and then weighed for bulk density determinations. Soil from the bulk density samples was portioned and used to measure particle-size distribution using a modified 12-hour hydrometer method (Bouyoucos, 1927), Mehlich-3 extractable soil nutrients (i.e., P, K, Ca, Mg, S, Na, Fe, Mn, Zn, and Cu; Mehlich, 1984; Tucker, 1992), soil pH, and electrical conductivity (EC).
Total water-stable aggregate (TWSA) concentration (i.e., > 0.25-mm diameter) was measured on samples collected from the 0-5 and 5-10 cm depths using a 4.8-cm diameter core chamber and slide hammer using a wet-sieving procedure (Yoder, 1936). Soil core samples were collected from areas in between switchgrass plants in switchgrass rows. Two replications were collected from each subplot. The two samples collected from each subplot were mixed together for one sample per depth interval. Each soil sample was manually broken down by hand and air-dried for 15 days. Approximately 400 g of air-air-dried soil was used for the wet-sieving procedure,
where soil was plunged in tap water at 30 cycles per minute for 5 minutes. Soil retained on the mesh openings of each of the 5 sieves (i.e., 4.0, 2.0, 1.0, 0.5, and 0.25 mm) was rinsed into an aluminum container with tap water. Water was decanted from the container, and remaining soil weighed after being dried at 70°C for 24 hours. Soil mass from each container was summed to calculate total mass of water stable aggregates from each subplot.
Soil samples for root density were collected from the 0-15 cm depth interval using a 7.3-cm diameter core chamber and slide hammer, and prepared according to the procedures followed by Brye and Riley (2009). Root samples were washed on a 2-mm mesh screen to collect the root material, dried at 55oC for 24 hours, and weighed. Root density core samples were collected from areas in between switchgrass plants in rows (interrow). One root core per subplot was collected in September 2013.
Double-ring infiltrometers were used to measure infiltration rates two days after a soaking rainfall. Moist antecedent soil moisture condition was intended to limit horizontal water movement into surrounding dry soil during infiltration measurements; infiltration measurements were intended to measure downward water movement in soil. Mature switchgrass was trimmed using hedge trimmers and residue carefully removed prior to placement of infiltrometers. One double ring infiltrometer measurement was conducted in each subplot in treatment combinations including cultivar, harvest frequency, and fertilizer source. In order to maximize potential infiltration differences among all other treatment combinations, irrigation measurements were only conducted in the non-irrigated plots. Double-ring infiltrometers were placed between switchgrass plants in switchgrass rows after switchgrass was mowed to a height of 6 cm in November 2013. Outer ring diameter was 30 cm, inner ring was 6 cm in diameter, and rings were 10 cm in height. The infiltrometer was inserted approximately 2 cm deep into the soil and outer
ring was filled with tap water to act as a buffer between dry soil and saturated soil in the inner ring. The inner ring was filled with tap water and the distance from the top of the soil to the water surface in the inner ring was measured at 0 minutes and after 1, 2, 3, 4, 5, 8, 10, 12, 15, 18, and 20 minutes.
Double-ring infiltration measurements were paired with mini-disk tension infiltrometer measurements (Decagon, Pullman, Washington) where infiltrometer tension was set at -2 cm in each plot. Two mini-disk infiltrometer measurements were conducted in each subplot, with one measurement collected from the center-ring area immediately following double ring infiltrometer measurements and one measurement collected in a nearby location. Two measurements were performed in each subplot for a total of 96 measurements. Both mini-disc and double-ring infiltrometers were placed between switchgrass rows (interrow), and plant bases were avoided to minimize leaking along root channels.
2.4. Statistical Analyses
The effects of cultivar, harvest frequency, irrigation, fertilizer source, soil depth, and their interactions on soil bulk density, SOM, soil chemical properties, particle-size distributions, and TWSA was evaluated by analysis of variance (ANOVA) using SAS (version 9.3, SAS Institute, Inc., Cary, NC). Similarly, an ANOVA was conducted to evaluate the effects of cultivar, harvest frequency, irrigation, fertilizer source, and their interactions on root density using SAS.
An analysis of covariance (ANCOVA) was conducted to evaluate the effects of cultivar, harvest frequency, irrigation, and fertilizer source on the relationship between infiltration rate
and time. When appropriate, means were separated by least significant difference (LSD) at α = 0.05.
3. Results
3.1. Soil Properties
3.1.1. Particle-size Distribution
Throughout the study site, sand, silt, and clay in the top 20 cm varied somewhat;
however, sand, silt and clay remained within the range of a silt-loam surface texture. Sand ranged from 19.8 to 36.8% and averaged 29.5%. Silt ranged from 42.9 to 55.8% and averaged 49.3%. Clay ranged from 14.2 to 29.2% and averaged 21.2%. However, sand, silt, and clay contents in the top 20 cm differed slightly among various treatment combinations, with the largest differences occurring between soil depths, which was expected (Table 1). Sand content differed between soil depths within cultivar-irrigation treatment combinations (p = 0.04; Table 1). In both the top 10 cm and in the 10-20 cm depth interval, sand content was greatest in the irrigated-‘Alamo’ treatment combinations, 32.6 and 30.3%, respectively (Table 2). Sand content in the top 10 cm and 10-20 cm depth interval was lowest in the non-irrigated-‘Cave-in-Rock’
(26.1%) and non-irrigated-‘Alamo’ (27.5%) treatment combinations (Table 2).
Averaged over all other treatment factors, silt content differed between soil depths (p <
0.01; Table 1), where silt content was greater in the top 10 cm than the 10- to -20 cm depth interval (50.4 and 48.2%, respectively). Silt (p= 0.04) and clay (p = 0.03) content also differed between fertilizer sources when averaged across all other treatment factors (Table 1), which was unexpected. Silt content was greater in the poultry litter (50%) than the commercial fertilizer
(48.6%) treatment combination. Clay content was greater in the commercial fertilizer (21.7%) than the poultry litter treatment combinations (20.6%). Averaged across all other treatment factors, silt content also differed between cultivars (p < 0.01; Table 1). Silt content was greater in
‘Cave-in-Rock’ (50.2%) than in ‘Alamo’ (48.4%) treatments. Despite the significant fertilizer source effect on silt and clay and significant cultivar effect on silt content, all silt and clay differences between treatments were less than 2%. Small differences of this magnitude are not likely to affect agronomic management.
Averaged over harvest frequencies, fertilizer sources, and irrigation treatments, clay content also differed between cultivars within soil depths (p < 0.01; Table 1). As expected, clay content increased with increasing depth from the top 10 cm (19.5%) to the 10- to -20 cm depth interval (22.9%) for both cultivars (Figure 4). Clay content was greatest and least for ‘Alamo’
treatment combinations in the 10- to -20 cm (22%) and 0-10 cm (17%) depth intervals, respectively.
The observed differences in sand, silt, and clay contents between soil depths were expected based on the reported textural classes of the top two horizons listed on the Leadvale silt-loam official series description (Appendix 1). Clay accumulation with increasing depth was expected, as the surface horizon (Ap) grades to an argillic (Bt) horizon at approximately 20 cm (Appendix 1). Despite some relatively minor differences in particle-size distribution among various treatment combinations, differences were not large enough to change the soil taxonomic classification anywhere in the study site and would likely not influence practical agronomic management decisions related to switchgrass production.
3.1.2. Bulk density
As expected, averaged over all other treatment factors, soil bulk density differed between soil depths (p < 0.001; Table 1). Soil bulk density was greater in the 10- to 20- cm depth (1.44 g cm-3) than in the top 10 cm (1.30 g cm-3).
Averaged over soil depths and cultivars, soil bulk density also differed between irrigation treatments within fertilizer-source-harvest-frequency treatment combinations (p = 0.004; Table 1). Soil bulk density was greater in the irrigated-poultry litter treatment combination with either harvest frequency (1.40 g cm-3) than the non-irrigated-1-cut treatment combination with either fertilizer source (1.33 g cm-3; Table 4). Soil bulk density was unaffected by switchgrass cultivar (p > 0.05; Table 1).
3.1.3. Aggregate Stability
Averaged over soil depth, cultivar, and harvest frequency, TWSA concentration differed between irrigation treatments within fertilizer sources (p = 0.03; Table 1). Total WSA
concentration was greater under irrigated (0.93 g g-1) than non-irrigated (0.86 g g-1) with poultry litter, but did not differ between irrigation treatments with commercial fertilizer (0.86 g g-1; Figure 4; Table 1).
Averaged over irrigation, TWSA concentration also differed between soil depths within cultivar-fertilizer-source-harvest-frequency treatment combinations (p = 0.02; Table 1). The greatest TWSA concentration was measured in the ‘Alamo’-commercial-fertilizer-2-cut
treatment combination in the 10- to -20 cm depth (0.93 g g-1), while the lowest TWSA
concentration was measured in the ‘Cave-in-Rock’-poultry-litter-1-cut treatment combination in the top 10 cm depth (0.87 g g-1; Table 3).
In addition, there were several notable trends in TWSA concentrations. Unexpectedly, TWSA concentration was generally numerically greater in ‘Alamo’ than in ‘Cave-in-Rock’
treatment combinations (Table 3). In contrast to hypothesized effects linking increased harvest frequency to decreased TWSA, TWSA concentration was generally numerically greater in the 2-cut than the 1-2-cut harvest frequency (Table 3). Total WSA concentration was generally
numerically greater in the 10- to -20 cm depth than in the top 10 cm (Table 3).
3.1.4. Soil pH and Electrical Conductivity
As expected, averaged over all other treatment factors, soil pH differed between fertilizer sources (p = 0.001; Table 7). Soil pH was greater in the poultry litter (pH = 6.1) than in the commercial fertilizer (pH = 5.9). Soil pH was unaffected by switchgrass cultivar, irrigation, harvest frequency, or soil depth. Though soil pH values differed between fertilizer sources, only 0.02 pH units separated the means.
Averaged over fertilizer source and soil depth, soil EC differed between cultivars within irrigation-harvest-frequency treatment combinations (p = 0.016; Table 7). Soil EC was greatest in ‘Cave-in-Rock’ (0.070 dS m-1) and least in ‘Alamo’ (0.056 dS m-1) treatments without irrigation in the 2-cut harvest frequency (Table 5).
Averaged over harvest frequency and fertilizer source, soil EC differed between soil depths within cultivar-irrigation treatment combinations (p = 0.006; Table 7). Soil EC was
generally greatest in the top 10 cm (Table 2). Soil EC was the lowest in the
non-irrigated-‘Alamo’ and irrigated-‘Cave-in-Rock’ treatment combinations in the 10- to -20 cm depth interval, which did not differ (Table 2).
Averaged over cultivar and irrigation, soil EC differed between soil depths within fertilizer-source-harvest-frequency treatment combinations (p = 0.020; Table 7). Soil EC was greatest in the 1-cut-poultry-litter treatment combination in the top 10 cm (0.081 dS m-1) compared to all other treatment combinations (Table 6). Soil EC did not differ among harvest-frequency-fertilizer-source treatment combinations in the 10- to -20 cm depth and averaged 0.058 dS m-1 (Table 6).
3.1.5. Extractable Soil Nutrients
Averaged over harvest frequency and fertilizer source, extractable soil Mn differed between soil depths within cultivar-irrigation treatment combinations (p = 0.026; Table 8).
Extractable soil Mn was greater in irrigated-‘Alamo’ (230.2 kg ha-1) treatment combination in the 10-20 cm depth interval than all other treatment combinations, which did not differ and averaged 176.6 kg ha-1 (Table 2).
Unexpectedly, extractable soil K (p = 0.033; Table 7) and Na (p = 0.019; Table 8) contents differed between cultivars when averaged over all other treatment factors. ‘Cave-in-Rock’ treatments had greater extractable soil K (95 kg K ha-1) and lower extractable soil Na (43 kg Na ha-1) compared to ‘Alamo’ (79 kg K and 64 kg Na ha-1) treatments.
Averaged over all other treatment factors, extractable soil Ca (p = 0.031; Table 7) and Fe (p < 0.001; Table 8) contents differed between soil depths. Extractable soil Ca was greater in the
10- to 20- cm depth (2079 kg ha-1) than in the top 10 cm (1965 kg Ca ha-1). In contrast,
extractable soil Fe was greater in the top 10 cm (216 kg Fe ha-1) than in the 0- to -10 cm depth (164 kg Fe ha-1).
Averaged over irrigation, cultivar, and harvest frequency, extractable soil P, K, Mg, S, Na, and Zn contents differed between soil depths within fertilizer-source treatments (p < 0.04;
Tables 7 and 8). The results for extractable soil P, K, Mg, S, and Zn contents followed similar trends related to treatment combinations and were generally greater in the top 10 cm compared to the 10- to -20 cm depth (Figures 4, 5, and 6). Extractable soil P, K, Mg, S, and Zn contents were greatest in the poultry-litter treatment combinations in the 0- to -10 cm depth and the lowest in the 10- to -20 cm depth treatment combinations of either fertilizer source, which did not differ (Figures 4, 5, and 6). In contrast, extractable soil Na was generally greater in the 10- to 20-cm depth compared to the top 10 cm, while extractable soil Na was greatest in the 10- to -20 cm-depth-poultry-litter and lowest in the 0- to -10 cm-commercial-fertilizer treatment combination (Figure 5).
Extractable soil P, K, Mg, and Zn contents differed between soil depths within harvest frequencies when averaged over irrigation, cultivar, and fertilizer source (p < 0.02; Tables 7 and 8). The results for extractable soil P, K, Mg, and Zn contents followed similar trends related to treatment combinations and were generally greater in the top 10 cm compared to the 10- to 20-cm depth (Figures 4, 5, and 6). Extractable soil P, K, Mg, and Zn contents were greatest in the 1-cut treatment combination in the top 10 cm, which supported the hypothesis that harvesting once after switchgrass senescence rather than twice per year allows greater retention of extractable soil nutrients (Figures 4, 5, and 6). Extractable soil P, K, Mg, and Zn contents were lower in the
10- to 20-cm than the 0- to 10-cm depth, but did not differ according to harvest frequency (Figures 4, 5, and 6).
In addition, when averaged over cultivar, harvest frequency, and soil depth, extractable soil Ca and Cu contents differed between fertilizer sources within irrigation treatment
combinations (p < 0.03; Tables 7 and 8). Extractable soil Ca and Cu contents followed similar trends, as both were the lowest in commercial-fertilizer and the greatest in irrigated-poultry-litter treatment combination (Figures 4 and 5).
Averaged over harvest frequency, fertilizer source, and soil depth, extractable soil P content differed between cultivars within irrigation treatments (p = 0.016; Table 7). Extractable soil P was lower in the irrigated-‘Cave-in-Rock’ than in ‘Alamo’ of either irrigation treatment and in the non-irrigated-‘Cave-in-Rock’ treatment combination, which did not differ (Figure 4).
Averaged over cultivar, harvest frequency, and fertilizer source, extractable soil Zn content differed between irrigation within soil depth treatments (p = 0.025; Table 8). While the greatest extractable Zn content was in the irrigated-0-to-10 cm depth compared to all other treatment combinations, extractable Zn content was generally greater in the top 10 cm than in the 10- to 20-cm depth (Figure 5). Extractable soil Zn was lowest in the 10- to 20-cm depth of either irrigation treatment compared to all other treatment combinations (Figure 5).
In addition, when averaged over soil depth, extractable soil Zn content differed between irrigation within cultivar-fertilizer-source treatments ((p = 0.43; Table 8). The greatest
extractable Zn content was in the irrigated-‘Alamo’-poultry-litter (2.3 kg ha-1) treatment, but did not differ from the ‘Cave-in-Rock’-poultry-litter treatment of either irrigation combination (Table 11). The lowest extractable Zn content was in the irrigated-‘Alamo’-commercial-fertilizer
(0.6 kg ha-1) treatment, but did not differ from all other commercial fertilizer treatments of either irrigation or cultivar treatment combination (Table 11).
The extractable soil Fe (p = 0.003) and Cu (p = 0.009) contents differed by cultivar within fertilizer-source-harvest-frequency treatment combinations when averaged over irrigation and soil depth (Table 8). Both extractable soil Fe and Cu were the greatest in the ‘Alamo’-poultry-litter-1-cut treatment combination and the lowest in the ‘Cave-in-Rock’-commercial-fertilizer-2-cut treatment combination compared to all other treatment combinations (Table 9).
Averaged over cultivar and irrigation, extractable soil Cu content differed by soil depth within harvest frequency-fertilizer-source treatment combinations (p = 0.038; Table 8).
Averaged over cultivar and irrigation, extractable soil Cu content differed by soil depth within harvest frequency-fertilizer-source treatment combinations (p = 0.038; Table 8).